Top 10 impacting data science use cases that could help you to solve business problems in supply chain management
Supply chain management (SCM) is the management of the flow of goods and services between businesses and locations that includes the movement and storage of raw materials. The objective of SCM is to maximize the service at the minimum cost level (the main objective is the satisfaction of the customers, of course).
Difficulties in SCM:
1)Uncertainty: Not sure how much stock of items will be needed tomorrow
2)Complexity of Data
3)High dependency on manual processes that leads to more data issues
4)Disruptions: e.g., Earthquakes in Japan & Hurricanes in the USA
I have been working for sometime now, trying to solve existing supply chain issues using data science solutions. I came across multiple stages of the supply chain, where I felt a data science solution would fairly improve the existing processes.
Hence, I would like to share a few use cases which could be implemented in the supply chain domain. Having examples at granule level would be difficult as every supply chain functions differently. Hence, in this article, we will see the top 3 use-case cases at levels which would be common across all SCM processes, i.e.
1. Supplier
2. DC
3. Destination/Outlet
3 use cases at each level mentioned above along with 1 use case which would be common across any level
I. Supplier
a) Vendor/Supplier Performance Management
Business use cases:
1. Assessment of Suppliers at DC side that helps higher management to monitor supplier performance
2. Identify the best suppliers as per their item supplying category using historical data. To assist in decision-making process in future for ordering goods.
3. Deploy a reinforcement learning supplier recommendation engine at DC receiver’s end. Each time goods are received; feedback of the goods delivery should be submitted. Next time, the model will recommend suppliers, considering input from the latest feedback.
b) Supplier Segmentation
Business use cases:
1) Cluster similar suppliers together in terms of types of goods they supply/ spend/ reliability/performance/delivery time etc.
2) Build a supplier recommendation system
c) Goods-Price contracts optimization
Business use cases:
1) Build a price recommendation system to identify best price of the goods supplied using mandatory parameters such as location, demand, current market value, seasonality, etc.
2) Build a ML model to identify optimal contract term
3) Build a system to identify best price hike for contract continuation
II. Distribution Center
a) Backorder Prediction
Business use cases:
1) Build a system to predict raw items that are going to be backordered or not
2) Optimize inventory management
3) Smooth deliveries of goods ordered
b) Lead Time Forecasting
Business use cases:
1) Build a system predict lead time of orders to avoid backorders and forecast the lead time for inventory replenishment
2) Inventory storage planning
3) Wastage analysis
c) Raw Items Case Quantities Received: Anomaly detection
Business use cases:
1) To avoid inconsistencies in inventory dashboards/reports, identify unexpected case quantities received. (The inconsistencies could have occurred due to any of the following reasons: fat fingering issue, duplicate entry of the good received at the DC, case quantity marked for the wrong raw item, etc.)
2) Optimized inventory planning and business decisions
3) Accurate pricing of the actual case quantity received.
III. Destination/Outlet
a) Dynamic Pricing
Business use cases:
1) Build a ML model to generate dynamic prices for products whose raw materials are about to expire.
2) Build a system to generate dynamic prices for products which are in high demand with limited supply
3) To improve revenue and minimize losses
b) Product Segmentation
Business use cases:
1) Segment different products (individual / category) based on sales, revenue, and risk/complexity, which could help higher management to take business decisions.
2) Get a product’s performance by understanding that products lie in which category
3) Build product recommendation system as per similar products
c) Staff forecasting & scheduling
Business use cases (This case could be applied at DC and outlet side on both levels, but it would be more effective at the outlet side as the footfall of the customer is highly inconsistent here throughout time):
1) Forecast the daily staff needed in the upcoming one/two weeks to ensure continuity in the service
2) Optimize utilization of resources/staff to improve operations efficiency and optimize resource cost
IV. Common Use Case
a) Invoice Data Extraction:
Business use cases:
1) Extract data from existing paper forms and have digital entries in the DB
2) Extract structure data from invoices so that, invoices can be automatically processed
3) Build an automated system to re-verify the data (related to payments, address, etc.) filled in the forms avoiding manual intervention
These are the top use cases along with possible business cases that could have a great impact on the current process if you are looking to implement data science solutions in supply chain management.
Please keep watching this space for more updates on these use cases & corresponding solutions. Thank you for reading; your feedback is highly appreciated.
